[openstack-dev] [TripleO] Auto-abandon bot

Ben Nemec openstack at nemebean.com
Wed Oct 14 14:41:25 UTC 2015

On 10/13/2015 08:08 PM, Dan Prince wrote:
> On Mon, 2015-10-12 at 11:38 -0500, Ben Nemec wrote:
>> On 10/10/2015 08:12 AM, Jeremy Stanley wrote:
>>> On 2015-10-09 17:10:15 -0500 (-0500), Ben Nemec wrote:
>>>> As discussed in the meeting a week or two ago, we would like to
>>>> bring
>>>> back the auto-abandon functionality for old, unloved gerrit
>>>> reviews.
>>> [...]
>>>> -WIP patches are never abandoned
>>> [...]
>>>> -Patches that are failing CI for over a month on the same patch
>>>> set
>>>> (regardless of any followup comments - the intent is that patches
>>>> expected to fail CI should be marked WIP).
>>> [...]
>>> Have you considered the possibility of switching stale changes to
>>> WIP instead of abandoning them?
>> That might be a valid alternative.  I'll bring it up in the meeting
>> tomorrow.
>>> I usually have somewhere around 50-100 open changes submitted
>>> (often
>>> more), and for some of those I might miss failures or negative
>>> review comments for a month or so at a time depending on what else
>>> I
>>> have going on. It's very easy to lose track of a change if someone
>>> abandons it for me.
>> I have to admit I had missed this point in previous discussions on
>> the
>> topic, and it is a valid concern IMHO.
>> As an alternative to just setting WIP on old changes, we could
>> significantly increase the abandon timeout.  I re-ran the tool with
>> some
>> different timeouts to get more data points.  Here's what I found:
>> Days - # of changes to be abandoned
>> 31 - 36
>> 90 - 29
>> 180 - 20
>> 365(!) - 12
> I gotta say the number of reviews we are talking about here doesn't
> seem to warrant a bot to me. Would it be reasonable to have a few core
> reviewers make a one time pass at the reviews in question and either
> abandon them (if appropriate) or mark them as WIP if the idea still
> stands but it needs more work?
> Perhaps we could revisit this again if the numbers above grow to say 75
> -100 on a regular basis, but I'm not sure we are at that level yet.

Honestly, to me any time that a core has to waste on a six month old
patch with unaddressed negative feedback is too much.  If the patch has
been sitting for six months and neither the committer nor another
reviewer can even be bothered to leave so much as a comment in that time
then it's clear to me the patch isn't important.

Basically what I'm saying is I'm not signing up to go through the list
of old patches and clean them up manually.  Are you?  If so, I'll wash
my hands of the whole issue, but I think _everyone_ has better things to do.

> Dan
>> So even setting the timeout to 1 year, which I think we can all agree
>> is
>> safe to call abandoned, we'd catch 12 changes.  Hence the desire for
>> a
>> cleanup bot. :-)
>> I kind of like 90 or 180 as a happy medium.  It would still clear a
>> non-trivial amount of patches, but I think after six months of
>> neglect
>> we can safely call a patch abandoned by the submitter.  I'm inclined
>> to
>> say 90 is safe too, but I'm okay with 180 and I do want to err on the
>> side of leaving things open so that would work for me.
>> _____________________________________________________________________
>> _____
>> OpenStack Development Mailing List (not for usage questions)
>> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubs
>> cribe
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

More information about the OpenStack-dev mailing list